AI Tools

hexstrike-ai

hexstrike-ai is an MCP server that bridges an AI assistant to a huge collection of security tools — roughly 150 of them, spanning reconnaissance, web testing, exploitation, and more. Once it’s connected to an MCP-capable assistant, you can describe what you want in plain language and the agent invokes the real tools to do it: run the scan, parse the output, decide the next step, and chain them together. It’s one of the clearest examples of agent-driven security work, and it composes with the rest of the toolkit rather than replacing it.

Oniomarchy sets hexstrike-ai up for you and puts it at Security → AI Tools → hexstrike-ai. You launch its server with hexstrike-server.

Using it

1. Start the server with hexstrike-server — it launches the MCP server that fronts the tool collection.
2. Connect an MCP client — point an MCP-capable AI assistant at the server so it can see and call the exposed tools.
3. Work in natural language — ask the agent to enumerate a host, probe a web app, or follow up on a finding, and it runs the underlying tools and reports back.

hexstrike-ai wraps the offensive tools; for driving Metasploit specifically through an agent, see metasploit-mcp, and for testing an LLM’s own defenses, see garak.